<scribe> scribe: JacekK
minutes from last week approved
Review of Action items [.1]. Editorial actions [.2]. ? 2004-04-01: Marsh will get schema tf going. ? 2004-09-02: Bijan to create stylesheet to generate a table of components and properties. ? 2004-09-16: Editors to move App C to RDF Mapping spec, except the frag-id which will move within media-type reg appendix. ? 2004-09-16: Editors to fix paragraph 6-9 of section 2.1.1 moved into 2.1.2 which talks about the syntax. ? 2004-10-14: Editors to add a statement like: The Style property may constrain both input and output, however a particular style may constrain in only one direction. In Section 126.96.36.199 of Part 1. (subsumed by LC21 resolution?) ? 2004-11-09: DaveO to work on text for option 3 (redefining conformance in terms of building the component model) (LC5f) ? 2004-11-09: DaveO will recast the @compatibleWith proposal using an extension namespace. (LC54) ? 2004-11-10: Sanjiva to write the rationale for rejecting LC75a ? 2004-11-10: Glen will post an e-mail describing the compromise proposal on formal objections. ? 2004-11-10: Editor remove ambiguity if it exists ? 2004-11-10: Sanjiva will write up this proposal and email it to the list as a response to the objection. ? 2004-11-11: Anish to propose additions to the test suite for the purpose of interoperability testing. ? 2004-11-11: Editors of part 2 and 3 to add text about WSDLMEP and SOAP mep mapping that points to section 2.3 of part 3 (LC48b) ? 2004-11-18: Mini-task force to propose one or two proposals for the group for LC5f. ? 2004-12-03: Glen and Asir to help craft the specfic text for the editors. ? 2004-12-03: Glen to send example on feature stuff for primer ? 2004-12-16: Part 3 Editors to update the HTTP binding with one of the above versions of text ? 2005-01-06: MTD Editors to add note saying content-type is not sufficient, information to be provided via other mechanism, for example xsi:type" ? 2005-01-06: MTD editors implement proposal 2 for issue 260. ? 2005-01-06: Umit? to respond to Larry, "not dynamic, other solutions equally bad, not recommendation track, if problems happy to consider those" ? 2005-01-13: Editors of media type doc to implement issue 261 resolution ? 2005-01-13: Editors of media type doc to implement issue 262 resolution ? 2005-01-13: Editors of media type doc to implement 262 and 273 ? 2005-01-13: Editors of media type doc to incorporate the text at 2004Dec/0022.html ? 2005-01-13: Editors of media type doc to resolve 275 editorially ? 2005-01-13: Umit to reply to issuer 270 ? 2005-01-13: Umit to respond to Ian Hickson about issue 271 ? 2005-01-19: Part 1 Editors to call out the difference between WSDL 1.1 and 2.0 in respect to single interface per service, and indicate alternatives ? 2005-01-19: Part 1 Editors to rewrite ONMR as Best practice. DONE [.4] 2005-01-19: Hugo to write a proposal for adding an optional Action attribute in line with WS Addressing (LC84b) ? 2005-01-19: Arthur and Asir to look for more edge cases ref LC20 and LC27. ? 2005-01-19: Part 1 Editors to add the intersection rule for f&p composition. ? 2005-01-20: Asir to think about mU and possibly propose some clarification text ? 2005-01-20: Arthur to come up with primer text to show fault reuse and fault code. DONE 2005-02-03: Marsh make Jan 0008 issue LC103 and close it. ? 2005-02-03: Members who have recently raised issues should check that these were reflected in the issues list. DONE [.6] 2005-02-03: Marsh to find schemaLocation supporting material for Arthur. DONE 2005-02-03: Marsh to add RDF mapping to agenda, try to recruit Jacek. ? 2005-02-03: Asir to work with primer eds on SOAP 1.1 example. ? 2005-02-03: Part 1 editors to incorporate text from Jan/0026 and Feb/0006. DONE [.3] 2005-02-10: Arthur to add non-normative ref to WS-Addr spec DONE [.5] 2005-02-10: Asir to say which component names need discussion Friendly amendment to original proposal. WG agrees to classify as editorial. ? 2005-02-10: DBooth to mail Arthur change to wording on media type registration, Arthur to incorporate. ? 2005-02-10: JMarsh to ask Henry to respond about media type issues with examples DONE 2005-02-10: Marsh to consider a cutoff for LC issues. Doesn't make sense with another LC release. DONE 2005-02-10: Marsh to move 106 up on the agenda. DONE 2005-02-10: Marsh to schedule discussion of LC70 along with LC5f. [.1] http://www.w3.org/2002/ws/desc/#actions [.2] http://www.w3.org/2002/ws/desc/4/lc-issues/actions.html [.3] http://lists.w3.org/Archives/Member/w3c-ws-desc/2005Feb/0022.html [.4] http://lists.w3.org/Archives/Public/www-ws-desc/2005Feb/0035.html [.5] http://lists.w3.org/Archives/Public/www-ws-desc/2005Feb/0041.html [.6] http://lists.w3.org/Archives/Public/www-ws-desc/2005Feb/0048.html
asir: my action done, friendly amendment to make the entire issue editorial
<asir> my amendment e-mail is at http://lists.w3.org/Archives/Public/www-ws-desc/2005Feb/0041.html
<Marsh> s me
Marsh: does anybody object to treating LC107 as editorial?
<scribe> ACTION: Marsh to reclasify LC107 as editorial and add reference to Asir's amendment [recorded in http://www.w3.org/2005/02/17-ws-desc-minutes#action01]
Marsh: cut-off for LC dismissed as we'll need a new LC anyway
jjm: I added table explaining how import and include actually work
<scribe> ACTION: Asir to review table on how import and include actually work (added by JJM) [recorded in http://www.w3.org/2005/02/17-ws-desc-minutes#action02]
ACTION 2=Asir to review table on how import and include actually work (added by JJM)
Marsh: don't have date for April,
but I'm looking at the week of 18th
... that would be a F2F joined with WS-Addressing, not co-hosted but close, same week
... Sunnyvale and Mountain View in Silicon valley
... next meeting probably May 30th, with Desc end of the week
Marsh: checking on status of progress, expectation of progress, scope of the draft
asir: can this work be done after WSDL 2?
Marsh: yes, alternatively I can report that this deliverable cannot be done at all
<dbooth> Jacek: I have interest in working on this. I've spent some time on this. Read draft ontology; probably will suggest changes. If only the ontology will be the content of the first public WD, then we might have it after the Tech Plenary.
<scribe> ACTION: jacekk to get RDF mapping work ongoing [recorded in http://www.w3.org/2005/02/17-ws-desc-minutes#action03]
bijan: was planning to do something concrete by next week, sending it to Jacek
Marsh: we also have a section on RDF mapping in the primer
dbooth: the intent was that this would be an awareness item in the primer, hi-level introduction about what the mapping is about
<scribe> ACTION: bijan and JacekK to supply text for primer section on RDF mapping [recorded in http://www.w3.org/2005/02/17-ws-desc-minutes#action04]
Marsh: security considerations section continues to be marked as "at risk"
Marsh: if we could have all editorial work done by the f2f, we'd only have the single issue
<Marsh> ACTION: Umit and Anish to complete editorial work on media type description before ftf [recorded in http://www.w3.org/2005/02/17-ws-desc-minutes#action05]
Umit: the media type document may actually precede the WSDL document
Marsh: we also need it for the MTOM recommendations
Marsh: 107 reclasified as
... some small issues on HTTP don't seem to be totally editorial
... objections to having 108, 109 reclasified as editorial?
... no objections, LC108 and LC109 are now editorial
Marsh: anything new?
Marsh: Arthur is not here today,
anybody objects to deferring this item until we have
... no objections
skipped because Umit doesn't have enough time on the call
Marsh: LC70 may be a precursor to LC5f, therefore now on the agenda
asir: part 1 allows multiple
schema languages, which we have concerns about
... 1) XML Schema is required, so decoupling from XML Schema is already hit
... 2) in the appendix on DTD and RelaxNG we seem to be inventing things to allow these languages; e.g. the DTD part puts a DTD inside a namespace
... for RelaxNG we redefine some of their semantics
... 3) it seems that mixed schema languages are allowed but it's not clear what happens when multiple languages are used
Marsh: what do you mean - simultaneous description of one type or just mixing?
... the question is - we don't know how they all come together
... 4) message reference has element attribute referring to element declaration but this doesn't indicate what schema language is in use
... so overloading of "element" AII is not fully clear
... two proposals:
... a) put XML Schema dependence out of part 1, indicate somehow what schema language is used
... b) drop RelaxNG and DTD from the spec, stay coupled to XML Schema
... additionally, if we have status quo or go with proposal a, we have to demonstrate implementations of use of other schema languages
... that's LC70
dbooth: regarding LC5f, option b is to treat xs:schema as an extension with implied wsdl:required="true"; does this route eliminate your concerns?
asir: seems like only concern 1
would go away
... we should talk about mixing within a class of extension - schema languages
alewis: I disagree with most of
... the appendices are non-normative, intended as examples, may be insufficient and that would have to be addressed, but not by dropping them
... we invent mechanisms in other places, it's inevitable, that's not a flaw
... no schema language suggests how it's going to be embedded or imported in other systems
... we may need clarifications, we may not need to suggest what happens with multiple schema languages used
... I don't see why anybody would actually use multiple schema langauges with conflicts; if they use a different type system, they probably don't want to use XML Schema at all
Roberto: agree 100% with amy
<Marsh> invention vs. illustration?
Roberto: going over the concerns - 1) is handled in LC5f, 2) we are inventing stuff in non-normative appendices, that's fine, provides an example
<pauld> is anyone working on an RDF Schema or OWL alternative to Schema in WSDL 2.0? Would such a thing make sense?
Roberto: 3) is a rathole
... if multiple type systems are understood, all components end up together as element declarations
... don't see the issues
asir: non-normative appendix of part 1 has a certain status, an example can go to the primer
Roberto: disagree it has any certain status
asir: one possibility to avoid
the rat hole is just allowing one schema system in one
... if schema systems are pluggable, XML Schema should be optional
<Zakim> JacekK, you wanted to suggest wsdxs:element, wsdldtd:element, wsdlrng:element and to suggest wsdxs:element, wsdldtd:element, wsdlrng:element; SOAP DM schema langauge
<dbooth> JacekK: 1. Asir wants an indication of what schema language is in use. It may seem weird that DTD and schema and RelaxNG might end up in the same bag, but ontology elements would not. Therefore maybe we could use different AIIs for different schema languages: One for XML Schema, one for RelaxNG, one for DTD.
<dbooth> ... 2. We have examples of equivalence to XML Schema. Maybe they could be referenced. 3. We could include OWL or SOAP Data Model Schema Language as an example of a situation where a totally different type system is used.
dorchard: agree with amy
... we do have the issue of potential conflicts; but dropping the support for other schema languages would be too drastic
... if other schema languages are in fact used, people will invent ways to do that and those can be mentioned in our errata or later publications
... we aren't losing anything if nobody uses this extensibility point, but if we limit it too much, we might lose out on a lot of innovation
<anish> q to ask if this can be a CR criterion
bijan: in favor of leaving the
... can we factor out the appendices?
Marsh: we'll try to segment the discussion down
<Zakim> alewis, you wanted to speak to issues of multiple definitions
<asir> i like bijan's proposal
alewis: reuse of the "element"
AII is a (mis)feature of the appendices specifically, not
mandated or forbidden by the extensibility mechanism
... we might have the various AIIs
... then our current element AII would be restricted solely to XML Schema
... we could rule out the same types or elements being described by multiple schema languages; but this seems to be getting us into problems
... ... of a processor being able to recognize such situations, but not knowing one of the languages used
... we might suggest a best practice, but not a robust way of detecting the problem
asir: point of information: XML Schema WG also talks about mixing schema languages
<asir> that issue is at http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC63
<dbooth> Jacek: Maybe if different AIIs were used, then we could limit to only one on each component.
JacekK: if we used the different AIIs, we could mandate the use of a single one on a message reference
alewis: we cannot have a MUST there, not testable
anish: what's the status of non-normative appendix and the normative extension for schema languages? do we need two impls of RelaxNG in CR?
alewis: not for a non-normative appendix
<asir> inventive, illustrative, speculative ..
Marsh: looks like disagreement on
the status of a non-normative appendix of the core spec or a
separate note or a section in the primer...
... seem to have same official status but give different perceptions
bijan: I thought a note had
easier chance of being a foundation for further work
... we could recommend to W3C that it give more work on the note but not on the other options
Marsh: I believe it would have the same effect
asir: I'd like to see a separate note
<Zakim> dbooth, you wanted to say But prohibiting the combination of schema languages would prohibit the use of a new schema language that is a superset of XML Schema.
dbooth: the appendix has two
different kinds of things - the beginning promises examples and
the rest contains MUSTs, that seems to be in discord for an
appendix, likely to cause confusion
... so anywhere but the appendices would be better
pauld: surprised to put this "work in progress" in part 1 appendix, better in primer or separately
alewis: don't understand dbooth's
criticism - it says "non-normative", if you're doing it, these
are the requirements
... I don't see a way of removing the potential confusion
dbooth: what you described is more like optional, not non-normative which can fully be ignored
<Roberto> +1 to Amy
dbooth: more disagreement on the status of non-normativeness
Marsh: perhaps a straw-poll is in
<Marsh> Option 0: status quo
<Marsh> Option 1: move to primer
<Marsh> Option 2: publish as note
<Marsh> Option 3: remove
<Marsh> Option 4: reword
jeffM: explained non-normative as an example, not something that can be conformed to
Marsh: straw-poll with multiple choice
Roberto: 0, 2
Marsh (MS hat on): 2,4
Marsh: option 0 has 6 votes
Marsh: option 1 has 4 votes
... option 2 has 12 votes
... option 3 has 2 votes
... option 4 has 7 votes
<bijan> The bijan option!
<asir> thanks bijan
Marsh: separate note seems to be
... no objections to not discussing removal of the section
... option 4 is like option 0, making it clearer that this is illustrative rather than optional
... are people who like 0 also comfortable with 4?
anish: me yes
Marsh: option 4 removes ambiguity
between non-normative and MUST
... more discussion of what "reword" means
<asir> yeh, illustrative and speculative
<bijan> This has to be the single most complex voting procedure/actual vote for, really, what is a minor issue!
Marsh: let's take this back to mailing list
<scribe> ACTION: Marsh to send out a poll today, allowing clarification of options by EOB tomorrow, poll next week before the call [recorded in http://www.w3.org/2005/02/17-ws-desc-minutes#action06]
<pauld> i've been using this as a reference model for an ircbot: http://stv.sourceforge.net/
Marsh: let's talk about the
"element" AII being ambiguous regarding what languages defines
... maybe we can have multiple attributes for different schema languages
... this would also make it clear in the case when there are duplicates between the schema languages
<Zakim> Roberto, you wanted to say something on this topic
JacekK: this would also help with incorrectly calling RelaxNG's element pattern as element declaration
Roberto: I'm for keeping the only
AII the way it is, when we designed it we knew what was
... we currently point to something that describes name, namespace name and content of an EII in infoset
Marsh: this imples multiples must be equivalent?
Roberto: we have a set of element declarations isomorphic to XML Schema
asir: we proposed leaving the element AII as it is but adding a selector attribute
JacekK: don't like the selector attribute because it may add redundancy in the case of totally different schema languages (like OWL)
<KevinL> If I remember right,a proposal was made by BEA to make the @element useable for other languages, but the group rejected that proposal and insisted the @element can only used to refer to XSD GED, other language has to define its own extensions. anybody remeber that?
Marsh: any more support for this change?
<alewis> KevinL, I recall exactly the reverse, that the proposal to forbid reuse was rejected.
Marsh: lack of response taken as lean towards status quo